# Voice attribution and redistribution receipt

The narrated edition was generated locally with Piper 1.4.2. Piper is used only as an
unvendored build tool and is not distributed with this repository.

- **Matt:** `en_GB-cori-medium`, `length_scale 1.08`, model SHA-256
  `1899f98e5fb8310154f3c2973f4b8a929ba7245e722b3d3a85680b833d95f10d`.
  [Model card](https://huggingface.co/rhasspy/piper-voices/blob/v1.0.0/en/en_GB/cori/medium/MODEL_CARD).
- **Ellie:** `en_US-kristin-medium`, `length_scale 0.98`, model SHA-256
  `5849957f929cbf720c258f8458692d6103fff2f0e3d3b19c8259474bb06a18d4`.
  [Model card](https://huggingface.co/rhasspy/piper-voices/blob/main/en/en_US/kristin/medium/MODEL_CARD).

Both model cards identify public-domain LibriVox source recordings. The Piper voice-model
repository is MIT-licensed. Piper 1.4.2 is GPL-3.0. Only generated narration and media,
this attribution, reproducibility source, and artifact hashes are published; neither
voice model nor the Piper engine is redistributed.

The earlier Microsoft Edge online-TTS narration is not reused. It remains withheld because
the free endpoint's client license did not establish redistribution rights for its output.